INDEX

Explanations

keywords typically used as formatting markers or instructions in a text editor

oai_token-act-pair · gpt-3.5-turbo

punctuation marks or symbols

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ichick

-0.72

inator

-0.71

stration

-0.71

unia

-0.71

 wisdom

-0.71

 fireplace

-0.70

pez

-0.70

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.70

rolet

-0.69

 extravag

-0.69

POSITIVE LOGITS

TON

0.92

AUT

0.87

âĢł

0.86

Thompson

0.82

 Shards

0.81

âĢ¢âĢ¢

0.77

Madison

0.77

.*

0.77

(*

0.76

Trade

0.76

Activations Density 0.017%

keywords typically used as formatting markers or instructions in a text editor

punctuation marks or symbols

No Comments

No Known Activations